No Backlight on LED TV

[Southerner1959]

I'm an electronics engineer, but I have limited experience of modern TVs. So I hope somebody can advise

I have an LG 47LN570V . . . which uses LEDs for the backlight.

When you turn it on, it flashes the LG Logo at full brilliance for about a quarter of a second . . . also when you turn it off. So firstly, I presume this means the LED arrays behind the screen must be OK?

Also, shining a torch on the screen you can see images, menus etc, and audio is working OK.

Despite downloading the Repair Manual, there are no circuit diagrams . . . so been trying to suss it out.

There are two separate feeds on the PSU board to the backlight . . . each with it's own diodes, smoothing caps and series Switching FET. I have replaced the capacitors . . . and measure 75V across each bank.

I have checked the voltage being fed to the Backlights . . . and there is just a quick pulse on switch on, in line with when the screen comes on for a moment.

So I guess it is the control circuitry going wrong, that presumably feeds a signal to the switching power FET, to adjust the intensity of the backlight?

I guess lots of possible faults could be causing this . . . so could I just bypass the FET, and feed the 75V to each LED feed? (it seems avery high voltage to me . . . . but I'm guessing there's a constant current regulator IC in each of the LED strings inside the screen?)

I know there would no longer be any Backlight adjustment . . . but this is my own set, and normally have it set at 100% anyway.

Any advice, or corrections to my assumptions would be much appreciated!

[fix2003]

its possible one of the led strips is faulty and shutting the circuit down for all  you will need to strip the screen down test each set see which one is faulty and Replace faulty LED strip

[downunder]

Suggest you also check out this Youtube video.

https://m.youtube.com/watch?v=Bf5T2wwtydY

[kennyc]

Watched the video . Puts you off even trying to fix the fault . He even has 3 faulty lights at the end of it .Even if you fix it and you give it back to the customer ,charge them  they will be back within a month with the same fault but just another LED gone. Thats if you dont break the screen while taking it apart. Remember as an apprentice years ago with Radio Rentals you did a screen clean on the old TVs and you would put it all back together and some dust would get on the tube and you had to do it all again 

[downunder]

Hi Kenny,

I agree, these videos usually provide more questions than answers. I only intended it as a guide to cracking open the panel, which would be too long to describe in detail. I open up maybe 2 a week and it does become quite routine. Ah well, back to rugby training.

Bruce

[Turnip]

Hi Kenny,

Used to do screen cleans on Pye VT4's where the plastic screen attracted even more dust - Mind you, it slipped out from the bottom so could suss it weekly.

Even worse were Cossor 950s where the tube became invisible.

Ex 'Visionhire' man myself where thinking folks bought thousands of 'Ferguson' and 'Peto Scott' tellies as they were cheap but total rubbish.

Admit Visionhire bought quite a lot of Philips sets, The mono versions were the biz, but the G500 series led to a decline as being quite the worst dual standard colour set 'designed' for somewhere else.

The G8 got it together again, but nothing near as reliable as anything else due to power and LOPT failures.

Sensible folks bought Japanese sets - Gather 'Visionhire' is now 'resting' at companies house, and has been for some years - Goes to show - Chris.

Hi Bruce - Didn't realise you were into Rugby - Guess passes the time - Chris.

[Southerner1959]

Thank you for the replies . . .

So are you suggesting that, although the screen looks fine for the split second it is on, it is detecting a fault in one of the LED strips and turning the voltage feed off?

BUT if there was a fault in one of the LED strips, why would it turn off BOTH of them? (each one is fed by a separate power rail, as mentioned above) - surely it would only cut the power to the faulty one?

PLUS - what could go wrong that would cause it to cut the power? Surely, if the LEDs fail, they go open circuit, and just don't work. (yet that can't be the case, as the whole screen looks fine for the split-second the Backlight flashes on)

I have checked all the voltages coming in and out of the PSU Board to the Main Board, and they are all correct.

I suspect there is just a fault in the Backlight dimming circuit on the PSU Board. If so, surely I can just by-pass the control circuit?

I gather that dimming of the backlight is achieved NOT by reducing the voltage, but by PWM, ie pulsing the feed and reducing the width of the pulses (similar to an ordinary room dimmer)

The Voltage going into each of the FETs is 75V . . . so presumably they are running at 100% at full brilliance, so I could just by-pass the FET on each feed and connect the 75V straight to each backlight strip?

But is there likely to be a constant-current driver circuit INSIDE the screen (attached to the LED Strings) . . . or is the power supply board feed likely to perform that function too? (I wouldn't have thought so)

[fix2003]

if one strip has a fault the monitoring circuit would normally shut them all down

[downunder]

Yes, fix2003 is correct. "2 seconds to black" is typically a backlighting problem. You have stated you have audio and can make out data on the screen. This means the TV is fully operational except for the backlights.

The earth return line for all the LED strips is monitored, and here any anomalies in the backlight are detected, an error signal is generated, and the backlights are extinguished. This cannot be confined to individual LED strips. One LED out, all LEDs out.

The duration of the brief flash you see is the time it takes for the protection to activate. If there is an ERROR
pin on the LED Driver Board, you'll note it sits at near 0V at switch on, then jumps up quickly as the error occurs.

Only sure way to know is to strip the panel and observe the LEDs at switch-on, to see which ones are inert.
You may not be comfortable with this as the display sheet is wafer-thin and easily cracked.

Bruce

[Southerner1959]

What I don't understand is HOW the backlights could come on for a moment and work perfectly on switch on and switch off if there was a fault?

What kind of fault with the LED strip could cause that?

Having said that, I have tried measuring the resistance of each of the strips, and can see nothing, is that normal?

[downunder]

They only appear to work perfectly, but if one strip is out, you're not going to pick it in the brief time you have, and the diffusion sheets behind the screen tend to equalize the brightness.
As I explained in my last post, when the TV is activated, whatever LED strips are operational will ignite, but those that don't will be detected as a current (I) imbalance and create an error signal to kill the backlights. This wont be instantaneous, hence the brief flash.

What can cause this? An open circuit LED, a shorted LED, faulty joiners (midway) in a LED strip, in fact any defect in the backlighting system. The protect circuit in the common earth return for the LEDs has the capacity to detect it.

AS for measuring the resistance of the strips, you'll get no measurement...that's normal. You can measure individual LEDs like a diode, and your meter voltage may even light the LED, if you can access the terminals.

Bruce

[Southerner1959]

Hi Bruce

Sorry, your previous reply was sent while I was writing mine, so hadn't seen that!

Thank you for explaining . . . I can see that makes sense now. The Power board is sending a quick pulse to the LEDs (I can see that on my meter) but turns them off straight away, so guess it is sensing something is wrong. (must be pretty smart if it can detect one LED is out)

I did try by-passing the FET that controls the feed to ONE of the two strips . . . . I fed it from the 75V supply, but through a 35 ohm resistor (for safety) . . .  but the lights didn't come on.

Guess the resistor was too high - the LEDs didn't take any current (ie there was still 75V on the output side of the resistor. Could I get away with feeding it direct from the supply, or does the dimming control circuit also provide current limiting?

EDIT: Just watched the Video you posted the Link to . . . the boards in that set look exactly the same as mine.  So that also confirms this must be the problem.

Sorry to everyone that replied on here . . . I just couldn't see why the picture was fine momentarily if some LEDs had failed . . . but guess the current-sensing must be pretty sensitive!

Looks one hell of a job to get it apart though!  (wonder how many that guy broke before he got the technique sussed?!)